[go: up one dir, main page]

EP4368108A1 - Procédés d'obtention d'un score de qualité - Google Patents

Procédés d'obtention d'un score de qualité Download PDF

Info

Publication number
EP4368108A1
EP4368108A1 EP22206530.2A EP22206530A EP4368108A1 EP 4368108 A1 EP4368108 A1 EP 4368108A1 EP 22206530 A EP22206530 A EP 22206530A EP 4368108 A1 EP4368108 A1 EP 4368108A1
Authority
EP
European Patent Office
Prior art keywords
metrics
obtaining
electrophysiological signals
electrophysiological
standardized
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP22206530.2A
Other languages
German (de)
English (en)
Inventor
Wadda DU TOIT
Cao Tri DO
Severine GISIN
Simon Bachmann
Moritz THIELEN
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Idun Technologies Ag
Original Assignee
Idun Technologies Ag
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Idun Technologies Ag filed Critical Idun Technologies Ag
Priority to EP22206530.2A priority Critical patent/EP4368108A1/fr
Publication of EP4368108A1 publication Critical patent/EP4368108A1/fr
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7221—Determining signal validity, reliability or quality
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/389—Electromyography [EMG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/398—Electrooculography [EOG], e.g. detecting nystagmus; Electroretinography [ERG]

Definitions

  • the human body is producing electrophysiological signals due to heart (ECG), muscle (EMG), eye (EOG) or brain (EEG) activity. These electric potentials can be measured using ExG recording devices, e.g. in-ear EEG devices using electrodes to measure the respective signals.
  • ECG heart
  • EEG muscle
  • EEG eye
  • EEG brain
  • EEG recording devices e.g. in-ear EEG devices using electrodes to measure the respective signals.
  • a metric is necessary to objectively compare and quantify the performance of the ExG data acquisition.
  • the current state-of-the-art solutions for estimating the quality of ExG can be categorized into four methods.
  • the present disclosure relates to a method for obtaining a plurality of quality metrics for electrophysiological signals, the method including the following steps: obtaining a plurality of metrics from each of: a first predetermined dataset of electrophysiological signals, a second predetermined dataset of electrophysiological signals, and a third predetermined dataset of electrophysiological signals; obtaining a plurality of standardized metrics of the plurality of metrics of each of the first, second and third predetermined datasets of electrophysiological signals using mean and standard deviation values of the plurality of metrics from the first predetermined dataset of electrophysiological signals; and obtaining a plurality of quality metrics as a reduced set of metrics from the plurality of standardized metrics based on an overlap between a) the respective standardized metrics of the first predetermined dataset of electrophysiological signals and the second predetermined dataset of electrophysiological signals, and b) the respective standardized metrics of the first predetermined dataset of electrophysiological signals and the third predetermined dataset of electrophysiological signals.
  • the number of the plurality of metrics is 52, but the present disclosure is not limited to this number.
  • the number of the plurality of metrics can be any number depending on the requirements of the specific application.
  • the overlap between the respective metrics refers to the overlap between the different curves (standardized distributions) of the respective metrics.
  • each standardized metric has a certain distribution, wherein the respective overlap relates to the overlap of the respective distributions or the overlap of the area underneath the respective curves (overlap of the respective integrals of the respective curves).
  • the overlap represents an association between the respective metrics and/or an equivalence between the respective metrics and/or a similarity between the respective metrics.
  • the plurality of metrics from each of the first, second and third predetermined datasets of electrophysiological signals are extracted from epochs comprising at least 250 samples, preferably between 250 and 7500 samples.
  • the sampling rate of the first, second and third predetermined datasets is 250 samples per second. That is, with this sampling rate, the plurality of metrics from each of the first, second and third predetermined datasets of electrophysiological signals are extracted from 1 second epochs, preferably between 1 and 30 second epochs and most preferably 5 second epochs.
  • sampling rates are possible and the present disclosure is not limited to a sampling rate of 250 samples per second.
  • the epochs overlap each other by a predetermined amount, wherein preferably a current evaluated epoch comprises 20% to 100% of new data compared to a previous epoch.
  • the overlap of epochs is used to have a large buffer while still being able to update the quality score with a low latency.
  • the overlap is in particular preferable for low latency updates.
  • any overlap is possible (even no overlap) as long as two conditions are satisfied: the current epoch consists of between 20% to 100% of new data and the current epoch has more than 250 samples.
  • the first predetermined dataset of electrophysiological signals corresponds to signals obtained by clinical gold standard electrode placement, e.g. 10-20 system for EEG, 12-lead precordial lead placement for ECG and the like.
  • the first predetermined dataset of electrophysiological signals corresponds to full scalp data, preferably measured at channel T4, and preferably good quality full scalp data.
  • the present disclosure is not limited to full scalp data and any dataset of electrophysiological signals that is known to represent good quality electrophysiological signals may be used, preferably from a gold standard database.
  • the first predetermined dataset of electrophysiological signals preferably corresponds to data from a gold standard database, preferably obtained by clinical gold standard electrode placement, and more preferably the first predetermined dataset (20) of electrophysiological signals corresponds to full scalp data measured at channel T4, preferably good quality full scalp data.
  • good quality in relation to electrophysiological signals as used herein may be defined by a low impedance of the electrode-skin contact, which is a widely accepted measure of good quality data. Impedance is a measure of the frequency dependent on contact resistance between the electrode and the subject's skin. The higher the impedance of the electrode, the smaller the amplitude of the EEG signal.
  • good quality data is defined by an impedance of the electrode-skin contact below 10 k ⁇ , more preferably below 5 k ⁇ and most preferably around 100 ⁇ .
  • the present disclosure is not limited by the measure impedance to define good quality data. Rather, other metrics could also have preferred ranges and quantify good quality data, e.g. SNR >2, amplitude between x and y in ⁇ V, etc.
  • obtaining the plurality of metrics comprises obtaining the plurality of metrics for each of a plurality of epochs and/or a plurality of datasets.
  • the second predetermined dataset of electrophysiological signals correspond to in-ear data, preferably good quality in-ear data
  • the third predetermined dataset of electrophysiological signals correspond to in-ear data, preferably bad quality in-ear data.
  • in-ear data refers to ExG data acquired from in-ear electrodes.
  • the second predetermined dataset of electrophysiological signals and/or the third predetermined dataset of electrophysiological signals correspond to ECG, EMG, or EOG signals and more preferably to EEG signals.
  • obtaining a plurality of quality metrics as a reduced set of metrics comprises: selecting a number of metrics with the largest overlap between the respective standardized metric of the first predetermined dataset of electrophysiological signals and the second predetermined dataset of electrophysiological signals, and/or selecting a number of metrics with the smallest overlap between the respective standardized metric of the first predetermined dataset of electrophysiological signals and the third predetermined dataset of electrophysiological signals.
  • the respective metric X refers to the raw value of the respective metric that is extracted from the current epoch.
  • X may refer to the alpha power from a 5 second epoch or the average signal amplitude of a 1 second epoch or the like.
  • any of the features/definitions provided in connection with following method for obtaining a quality score percentage function for electrophysiological signals and the method for obtaining a quality score for an electrophysiological signal apply to the method for obtaining a plurality of quality metrics for electrophysiological signals accordingly.
  • the present disclosure also relates to a method for obtaining a quality score percentage function for electrophysiological signals, the method including the following steps: obtaining a plurality of metrics from each of: a first predetermined dataset of electrophysiological signals, a second predetermined dataset of electrophysiological signals, and a third predetermined dataset of electrophysiological signals; obtaining a plurality of standardized metrics of the plurality of metrics of each of the first, second and third predetermined datasets of electrophysiological signals using mean and standard deviation values of the plurality of metrics from the first predetermined dataset of electrophysiological signals; obtaining a weighting parameter W for each of the plurality of standardized metrics based on an overlap between respective standardized metrics of the first predetermined dataset of electrophysiological signals and the second predetermined dataset of electrophysiological signals, an overlap between respective standardized metrics of the first predetermined dataset of electrophysiological signals and the third predetermined dataset of electrophysiological signals, and an area of the respective standardized metrics of the first predetermined dataset of electrophysiological signals; obtaining a plurality of weighted standardized metrics
  • the plurality of metrics correspond to a plurality of quality metrics obtained by a method according to the above described method for obtaining a plurality of quality metrics for electrophysiological signals.
  • the plurality of quality metrics may also be chosen by a different method or based on prior knowledge of the metrics. That is, the method for obtaining a quality score percentage function for electrophysiological signals is not limited to be used in combination with the above described method for obtaining a plurality of quality metrics for electrophysiological signals.
  • the respective metric X refers to the raw value of the respective metric that is extracted from the current epoch.
  • X may refer to the alpha power from a 5 second epoch or the average signal amplitude of a 1 second epoch or the like.
  • the plurality of weighted standardized metrics are obtained by multiplying the plurality of standardized metrics with the respective weighting parameter.
  • obtaining the plurality of metrics comprises obtaining the plurality of metrics for each of a plurality of epochs.
  • obtaining a quality score percentage function comprises summing the weighted standardized metrics for each of the plurality of epochs.
  • obtaining a quality score percentage function comprises designating predetermined percentages to each of the plurality of summed weighted standardized metrics, wherein preferably the lowest summed weighted standardized metric is designated 100% and the highest summed weighted standardized metric is designated 50%.
  • obtaining a quality score percentage function comprises fitting the designated predetermined percentages to each of the plurality of summed plurality of weighted standardized metrics, wherein the fitted function corresponds to the quality score percentage function.
  • the present disclosure also relates to a method for obtaining a quality score for an electrophysiological signal, the method including the following steps: obtaining a plurality of metrics from the electrophysiological signal; obtaining a plurality of standardized metrics of the plurality of metrics based on mean and standard deviation values of a plurality of respective metrics from a first predetermined dataset of electrophysiological signals; obtaining a weighting parameter for each of the plurality of standardized metrics based on an overlap between respective standardized metrics of the first predetermined dataset of electrophysiological signals and a second predetermined dataset of electrophysiological signals, an overlap between respective standardized metrics of the first predetermined dataset of electrophysiological signals and a third predetermined dataset of electrophysiological signals, and an area of the respective standardized metrics of the first predetermined dataset of electrophysiological signals; obtaining a plurality of weighted standardized metrics of the plurality of metrics from the electrophysiological signal based on the standardized metrics and the respective weighting parameter; and obtaining a quality score based on the plurality of weighted standardized metrics of
  • the plurality of metrics correspond to a plurality of quality metrics obtained by a method according to the above described method for obtaining a plurality of quality metrics for electrophysiological signals.
  • obtaining a quality score comprises applying a quality score percentage function, preferably obtained by the method for obtaining a quality score percentage function for electrophysiological signals as described above, to the plurality of weighted standardized metrics of the electrophysiological signal.
  • obtaining a quality score comprises obtaining an average of the plurality of weighted standardized metrics of the electrophysiological signal.
  • obtaining a quality score comprises obtaining a plurality of weighted standardized metrics of the electrophysiological signal for each of a plurality of epochs.
  • obtaining a quality score comprises obtaining a sum of the plurality of weighted standardized metrics of the electrophysiological signal for each of the plurality of epochs.
  • the method comprises obtaining a heatmap based on the plurality of weighted standardized metrics of the electrophysiological signal for each of a plurality of epochs and/or the sum of the plurality of weighted standardized metrics of the electrophysiological signal for each of the plurality of epochs.
  • obtaining a quality score comprises obtaining an average of the sum of the plurality of weighted standardized metrics for each of the plurality of epochs.
  • obtaining a quality score comprises obtaining an average of the sum of the quality scores for each of the plurality of epochs
  • any of the features/definitions provided above in connection with the method for obtaining a plurality of quality metrics for electrophysiological signals and/or the method for obtaining a quality score percentage function for electrophysiological signals apply to the method for obtaining a quality score for an electrophysiological signal.
  • the present disclosure also relates to a combination of the method for obtaining a plurality of quality metrics for electrophysiological signals and/or the method for obtaining a quality score percentage function for electrophysiological signals and/or the method for obtaining a quality score for an electrophysiological signal.
  • the present disclosure also relates to a computer-implemented method comprising the steps of the method for obtaining a plurality of quality metrics for electrophysiological signals and/or the method for obtaining a quality score percentage function for electrophysiological signals and/or the method for obtaining a quality score for an electrophysiological signal.
  • the present disclosure also relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method for obtaining a plurality of quality metrics for electrophysiological signals and/or the method for obtaining a quality score percentage function for electrophysiological signals and/or the method for obtaining a quality score for an electrophysiological signal.
  • the use of "obtaining” may encompass “calculating” or “processing”, e.g. by the respective unit, component or processor, and “receiving”, e.g. from a different unit, component or processor.
  • the methods disclosed above are thus simple, data driven and can be used in real-time.
  • the methods do not require prior knowledge about which features are indicative of a good ExG signal and require no knowledge on which thresholds to set.
  • the methods are based on resting state data and do not require a stimulation or expected results from the stimulation.
  • the methods do not require any estimations of noise and are thus suitable for single channel ExG signals.
  • EEG is chosen as an example of a specific use case for quality assessment, but this methodology can be used on any ExG signal where a gold standard database, i.e. predetermined good data, is available.
  • Fickling, S.D. et al. the goals and algorithms of the present disclosure are fundamentally different.
  • Fickling, S.D. et al.'s similar component is used for automatic noise rejection by classifying good and bad short time frames in the time-series signal (called 'epochs').
  • the goal of the present disclosure's use of a similar component is to evaluate and estimate a quality score for each epoch and entire datasets, which can be used to quantify the overall quality of the signal measurements. This is to aid development of ExG signal measurement devices as well as aid in the analysis of ExG data in terms of measurement capabilities.
  • Fickling, S.D. et al. use metrics based on prior knowledge.
  • the methods of the present disclosure provide a pipeline, which eliminates the need for prior knowledge about which metrics to use to evaluate an ExG signal. This makes it possible to create a quality score for any ExG signal where a gold standard database is available.
  • Fickling, S.D. et al. weigh all the metrics with the same importance. According to the present disclosure the importance of each metric is quantified and incorporated into the quality score. These weights are based on how similar good ExG signals are from the gold standard database and how different bad ExG signals are from the gold standard database.
  • Fickling, S.D. et al. uses the standard difference to indicate bad and good epochs, whereas the present disclosure adds weighting to the calculation and uses a novel methodology for evaluating and comparing ExG signals to a gold standard database to inform successful or unsuccessful development.
  • Fickling, S.D. et al. uses a quality indication by summing the standard differences of theoretically informed metrics and using the amount of total zero standard differences as an indication of bad data.
  • the present disclosure uses a pipeline that automatically, data driven, identifies the best metrics to use and weighs these metrics according to importance and only then sums the weighted standard differences.
  • the present disclosure defines quality score percentages according to the distribution of the summed weighted differences. This allows to create a function that converted weighted summed differences to a quality score between 0 and a 100%
  • the present disclosure is not limited to the exemplary embodiments and applications described and illustrated herein. Additionally, the specific order and/or hierarchy of steps in the methods disclosed herein are merely exemplary approaches. Based upon design preferences, the specific order or hierarchy of steps of the disclosed methods or processes can be re-arranged while remaining within the scope of the present disclosure. Thus, those of ordinary skill in the art will understand that the methods and techniques disclosed herein present various steps or acts in a sample order, and the present disclosure is not limited to the specific order or hierarchy presented unless expressly stated otherwise.
  • Fig. 1 shows an exemplary EEG signal 10.
  • Fig. 1 also illustrates the use of the Z-score Zto standardize the metrics.
  • Fig. 1 illustrates the overlap of the different standardized metrics for identifying suitable quality metrics.
  • a plurality of metrics are extracted from 5 second epochs 11 for each dataset (i.e. first dataset 20, second dataset 21 and third dataset 22).
  • the black square in the exemplary EEG signal 10 illustrates a 5 second epoch 11.
  • the 5 second epochs 11 overlap each other by 4 seconds.
  • the present disclosure is not limited to the specific values for the size of the epochs 11 and the overlap.
  • the size of the epochs 11 and the overlap are preferably chosen according to the sample rate.
  • the exemplary 5 second epoch 11 would contain 1250 samples. In principle, anything above 250 samples may be sufficient for the method as disclosed herein.
  • the purpose of the overlap is to have a large buffer while still being able to update the quality score with a low latency.
  • the overlap is thus only necessary for low latency updates. Therefore, 0% overlap is also possible, as long as the following epoch is longer than 250 samples. Therefore, any overlap is possible as long as two conditions are satisfied: the current evaluated epoch consists of between 20% to 100% of new data and the current epoch has > 250 samples.
  • the good quality full scalp data (first dataset 20) are preferably taken from a gold standard database.
  • the data may be quantified as good data by reference to the impedance of the electrode-skin contact, which is a widely accepted measure of good quality data.
  • the higher the impedance of the electrode-skin contact the smaller the amplitude of the EEG signal 10 may be.
  • impedance smaller than 200 kOhm may be acceptable, below 20 kOhm may be good, and below 10 kOhm may be preferable.
  • other metrics could also have preferred ranges, e.g. SNR >2, amplitude between x and y in ⁇ V, etc.
  • the Z-scores Z of the metrics are calculated for each epoch 11 and each dataset using the mean and standard deviation values of the good full scalp data 20 for standardization.
  • good quality in-ear EEG data (second dataset 21) metrics are standardized using the mean and standard deviation of the respective metrics of the good quality full scalp data 20.
  • bad quality EEG data (second dataset 22) metrics are standardized using the mean and standard deviation of the respective metrics of the good quality full scalp data 20.
  • the 52 metrics are then filtered down to only relevant metrics using the following method.
  • the respective metrics (distributions thereof) of good quality full scalp data 20, good quality in-ear EEG data 21 and bad quality in-ear EEG data 22 are illustrated at the bottom of Fig. 1 .
  • the area below the curve of the good quality full scalp data 20 is indicated by C.
  • the overlap between the area below the curve of the good quality full scalp data 20 and the good quality in-ear EEG data 21 is indicated by A and the overlap between the area below the curve of the good quality full scalp data 20 and the bad quality in-ear EEG data 22 is indicated by B.
  • a weighted Z-score is obtained as Z ⁇ W.
  • the respective weighted Z-scores are then calculated only for the most relevant metrics (according to the above assumptions) for good quality full scalp data 20, good quality in-ear EEG data 21, and bad quality in-ear EEG data 22 for all epochs 11.
  • the weighted Z-scores are summed for each epoch. It should be noted that this step is mainly for evaluation purposes. That is, this step of calculating the respective weighted Z-scores for good quality full scalp data 20, good quality in-ear EEG data 21, and bad quality in-ear EEG data 22 for all epochs 11 is mainly to show that the method of identifying the most relevant metrics performs well. In a real scenario of obtaining a quality score for a measured signal of unknown quality, the identification of the relevant metrics and the calculation of the weighting parameter is not performed with the measured signal, but performed prior to evaluating the measured signal (as outlined in more detail below).
  • Fig. 2 shows the summed and weighted Z-scores for the good quality full scalp data 20, the good quality in-ear EEG data 21 and the bad quality in-ear EEG data 22.
  • the metrics chosen provide a good correlation between good quality full scalp data 20 and good quality in-ear EEG data 21 and also differentiates good from bad data 22.
  • Fig. 3 shows an exemplary EEG signal 10, wherein an epoch 11 is again indicated by a black square.
  • the weighted Z-scores of the relevant metrics chosen are calculated as described with reference to Figs. 1 and 2 for each epoch 11 and each metric for the electrophysiological signal 10, i.e. the measured in-ear EEG signal of unknown quality using the mean and standard deviation values of the good quality full scalp data 20. That is, the relevant metrics for in-ear EEG signals are determined in accordance with the above described method, i.e. with reference to good quality full scalp data 20, good quality in-ear EEG data 21 and bad quality in-ear EEG data 22 (all of which are pre-characterized as such). The respective metrics are then calculated from the measured in-ear EEG signal of unknown quality using the mean and standard deviation values of the good quality full scalp data 20.
  • weighting parameters are determined as described above for the determined relevant metrics from the, i.e. with reference to good quality full scalp data 20, good quality in-ear EEG data 21 and bad quality in-ear EEG data 22, and multiplied with the Z-scores of the measured in-ear EEG signal of unknown quality to obtain the weighted Z-scores of the measured in-ear EEG signal of unknown quality.
  • the respective Z-scores can be arranged in a matrix structure as illustrated in Fig. 3 . That is, each row may be assigned to a specific metric 1 to N and each column may be assigned to a specific epoch 1 to N. The last row of the matrix may represent the summed weighted Z-scores of each column.
  • the matrix representation of Fig. 3 may also be visualized by a heatmap, i.e. a color-coded matrix for easy interpretation of the signal under investigation.
  • the summed and weighted Z-scores may be matched with percentages based on the experience of what is considered a good and what is considered a bad Z-score. For example, the lowest summed weighted standardized metric is designated 100% and the highest summed weighted standardized metric is designated 50%. The outliers above that completed the range between 50% and 0%.
  • the respective Z-scores with their corresponding x and y values are plotted and a conversion function is fitted through it.
  • the Z-scores are then multiplied by the function shown in Fig. 5 to obtain a percentage value for each epoch. From this, a final percentage score based on the average percentage for all the epoch scores may be obtained as a quality score of the overall signal.
  • the first part of the present disclosure is finding the optimal metrics that differentiates good ExG data from bad ExG data. These metrics should be in comparable ranges between good full-scalp EEG data and good in-ear EEG data. 52 metrics were extracted from good full-scalp EEG data at the location T4. T4 is a standard EEG electrode location from the 10-20 system of electrode placement and is in close proximity to the right ear. Therefore, the signals compare to in-ear EEG signals.
  • the standard Z-scores are calculated for good and bad in-ear EEG by using the gold-standard T4 data as a reference. This means the mean and standard deviation used in the standardisation is always based on the gold-standard database. Standardisation in this context means that the metrics from the in-ear EEG are re-calculated by calculating their corresponding Z-score values as shown in the formula for the standard score. In other words, the mean and standard deviation of the good full scalp metrics is used to standardize the good quality in-ear EEG data metrics and the bad quality in-ear EEG metrics.
  • the metrics that were the most similar in terms of distribution between the standardized full-scalp data and standardized good in-ear data and most different between standardized full-scalp data and the standardized bad in-ear EEG data is used as the final quality metrics (based on the overlap of the curves). That is, the final quality metrics are used as a subset of the 52 metrics.
  • the second part of the present disclosure is to create a score from the quality metrics.
  • Weighted Z-score values are calculated for the full-scalp data. The lowest weighted Z-score values from the full scalp data is used as a quality score of 100%. The highest weighted Z-score values is used as 50%. The outliers above that completed the range between 50% and 0%. From the full scalp weighted Z-score values the percentages is used as the y-axis and the weighted Z-score values are used as the x-axis and a conversion from Z-score value to quality score percentage function was created.
  • the third part is the application of the above mentioned methods to obtain a quality score for EEG data.
  • a 5 second epoch is extracted from the time-domain data.
  • the relevant metrics are calculated.
  • the metrics are standardized using the mean and standard deviation from the good full scalp data.
  • the metrics are then multiplied by a weight based on how well the metrics are at differentiating good from bad data as well as how similar are the good in-ear EEG data and the full scalp data for that metric. All the weighted Z-score metrics are then summed to have a final epoch value. This value is then converted to a quality percentage score using the fitted function.
  • the method is then repeated for each overlapping epoch of the dataset and finally all the scores of each epoch is averaged to obtain a score of the complete dataset.
  • the method may be used as real-time quality assessment, i.e. the method may also be applied to only one epoch and there is no need to analyse the final result over all epochs.
  • Fig. 6 shows a schematic diagram of a method according to an embodiment of the present disclosure.
  • the method may be used for obtaining a plurality of quality metrics for electrophysiological signals.
  • the method includes the following steps:
  • the plurality of metrics from each of the first, second and third predetermined datasets of electrophysiological signals are extracted from epochs comprising at least 250 samples, preferably between 250 and 7500 samples.
  • the epochs overlap each other by a predetermined amount, wherein preferably a current evaluated epoch comprises 20% to 100% of new data compared to a previous epoch.
  • the first predetermined dataset of electrophysiological signals corresponds to signals obtained by clinical gold standard electrode placement, e.g. 10-20 system for EEG, 12-lead precordial lead placement for ECG and the like.
  • the first predetermined dataset of electrophysiological signals corresponds to full scalp data measured at channel T4, preferably good quality full scalp data.
  • the first predetermined dataset of electrophysiological signals preferably corresponds to data from a gold standard database, preferably obtained by clinical gold standard electrode placement, and more preferably the first predetermined dataset (20) of electrophysiological signals corresponds to full scalp data measured at channel T4, preferably good quality full scalp data
  • obtaining the plurality of metrics comprises obtaining the plurality of metrics for each of a plurality of epochs.
  • the second predetermined dataset of electrophysiological signals correspond to in-ear data, preferably good quality in-ear data, and/or wherein the third predetermined dataset of electrophysiological signals correspond to in-ear data, preferably bad quality in-ear data.
  • obtaining a plurality of quality metrics as a reduced set of metrics comprises: selecting a number of metrics with the largest overlap between the respective standardized metric of the first predetermined dataset of electrophysiological signals and the second predetermined dataset of electrophysiological signals, and/or selecting a number of metrics with the smallest overlap between the respective standardized metric of the first predetermined dataset of electrophysiological signals and the third predetermined dataset of electrophysiological signals.
  • the respective metric X refers to the raw value of the respective metric that is extracted from the current epoch.
  • X may refer to the alpha power from a 5 second epoch or the average signal amplitude of a 1 second epoch or the like.
  • Fig. 7 shows a schematic diagram of a method according to an embodiment of the present disclosure.
  • the method may be used for obtaining a quality score percentage function for electrophysiological signals.
  • the method includes the following steps:
  • the plurality of metrics correspond to a plurality of quality metrics obtained by the method for obtaining a plurality of quality metrics for electrophysiological signals as described above.
  • the respective metric X refers to the raw value of the respective metric that is extracted from the current epoch.
  • X may refer to the alpha power from a 5 second epoch or the average signal amplitude of a 1 second epoch or the like.
  • the plurality of weighted standardized metrics are obtained by multiplying the plurality of standardized metrics with the respective weighting parameter.
  • obtaining the plurality of metrics comprises obtaining the plurality of metrics for each of a plurality of epochs.
  • obtaining a quality score percentage function comprises summing the weighted standardized metrics for each of the plurality of epochs.
  • obtaining a quality score percentage function comprises designating predetermined percentages to each of the plurality of summed weighted standardized metrics, wherein preferably the lowest summed weighted standardized metric is designated 100% and the highest summed weighted standardized metric is designated 50%.
  • obtaining a quality score percentage function comprises fitting the designated predetermined percentages to each of the plurality of summed plurality of weighted standardized metrics, wherein the fitted function corresponds to the quality score percentage function.
  • Fig. 8 shows a schematic diagram of a method according to an embodiment of the present disclosure.
  • the method may be used for obtaining a quality score for an electrophysiological signal.
  • the method includes the following steps:
  • the plurality of metrics correspond to a plurality of quality metrics obtained by the method for obtaining a plurality of quality metrics for electrophysiological signals as described above.
  • obtaining a quality score comprises applying a quality score percentage function, preferably obtained by the method for obtaining a quality score percentage function for electrophysiological signals as described above, to the plurality of weighted standardized metrics of the electrophysiological signal.
  • obtaining a quality score comprises obtaining an average of the plurality of weighted standardized metrics of the electrophysiological signal.
  • obtaining a quality score comprises obtaining a plurality of weighted standardized metrics of the electrophysiological signal for each of a plurality of epochs.
  • obtaining a quality score comprises obtaining a sum of the plurality of weighted standardized metrics of the electrophysiological signal for each of the plurality of epochs.
  • the method comprises obtaining a heatmap based on the plurality of weighted standardized metrics of the electrophysiological signal for each of a plurality of epochs and/or the sum of the plurality of weighted standardized metrics of the electrophysiological signal for each of the plurality of epochs.
  • obtaining a quality score comprises obtaining an average of the sum of the plurality of weighted standardized metrics for each of the plurality of epochs.
  • a computer-implemented method comprising the steps of the method for obtaining a plurality of quality metrics for electrophysiological signals and/or the method for obtaining a quality score percentage function for electrophysiological signals and/or the method for obtaining a quality score for an electrophysiological signal is provided.
  • a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method for obtaining a plurality of quality metrics for electrophysiological signals and/or the method for obtaining a quality score percentage function for electrophysiological signals and/or the method for obtaining a quality score for an electrophysiological signal is provided.
  • any reference to an element herein using a designation such as "first,” “second,” and so forth does not generally limit the quantity or order of those elements. Rather, these designations can be used herein as a convenient means of distinguishing between two or more elements or instances of an element. Thus, a reference to first and second elements does not mean that only two elements can be employed, or that the first element must precede the second element in some manner.
  • any one of the various illustrative logical blocks, units, processors, means, circuits, methods and functions described in connection with the aspects disclosed herein can be implemented by electronic hardware (e.g., a digital implementation, an analog implementation, or a combination of the two), firmware, various forms of program or design code incorporating instructions (which can be referred to herein, for convenience, as "software” or a "software unit”), or any combination of these techniques.
  • a processor, device, component, circuit, structure, machine, unit, etc. can be configured to perform one or more of the functions described herein.
  • IC integrated circuit
  • DSP digital signal processor
  • ASIC application specific integrated circuit
  • FPGA field programmable gate array
  • the logical blocks, units, and circuits can further include antennas and/or transceivers to communicate with various components within the device.
  • a general purpose processor can be a microprocessor, but in the alternative, the processor can be any conventional processor, controller, or state machine.
  • a processor can also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other suitable configuration to perform the functions described herein. If implemented in software, the functions can be stored as one or more instructions or code on a computer-readable medium. Thus, the steps of a method or algorithm disclosed herein can be implemented as software stored on a computer-readable medium.
  • Computer-readable media includes both computer storage media and communication media including any medium that can be enabled to transfer a computer program or code from one place to another.
  • a storage media can be any available media that can be accessed by a computer.
  • such computer-readable media can include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired program code in the form of instructions or data structures and that can be accessed by a computer.
  • unit refers to software, firmware, hardware, and any combination of these elements for performing the associated functions described herein. Additionally, for purpose of discussion, the various units are described as discrete units; however, as would be apparent to one of ordinary skill in the art, two or more units may be combined to form a single unit that performs the associated functions according to embodiments of the present disclosure.
  • memory or other storage may be employed in embodiments of the present disclosure.
  • memory or other storage may be employed in embodiments of the present disclosure.
  • any suitable distribution of functionality between different functional units, processing logic elements or domains may be used without detracting from the present disclosure.
  • functionality illustrated to be performed by separate processing logic elements, or controllers may be performed by the same processing logic element, or controller.
  • references to specific functional units are only references to a suitable means for providing the described functionality, rather than indicative of a strict logical or physical structure or organization.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Medical Informatics (AREA)
  • Surgery (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Veterinary Medicine (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Psychiatry (AREA)
  • Molecular Biology (AREA)
  • Physics & Mathematics (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Psychology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Signal Processing (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
EP22206530.2A 2022-11-10 2022-11-10 Procédés d'obtention d'un score de qualité Withdrawn EP4368108A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP22206530.2A EP4368108A1 (fr) 2022-11-10 2022-11-10 Procédés d'obtention d'un score de qualité

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
EP22206530.2A EP4368108A1 (fr) 2022-11-10 2022-11-10 Procédés d'obtention d'un score de qualité

Publications (1)

Publication Number Publication Date
EP4368108A1 true EP4368108A1 (fr) 2024-05-15

Family

ID=84330994

Family Applications (1)

Application Number Title Priority Date Filing Date
EP22206530.2A Withdrawn EP4368108A1 (fr) 2022-11-10 2022-11-10 Procédés d'obtention d'un score de qualité

Country Status (1)

Country Link
EP (1) EP4368108A1 (fr)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3335630A1 (fr) * 2016-12-15 2018-06-20 Mensia Technologies Indice de qualité de signal améliorée de bio-signal multicanal utilisant la géométrie de riemann

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3335630A1 (fr) * 2016-12-15 2018-06-20 Mensia Technologies Indice de qualité de signal améliorée de bio-signal multicanal utilisant la géométrie de riemann
WO2018109166A1 (fr) * 2016-12-15 2018-06-21 Mensia Technologies Amélioration de l'indice de qualité du signal d'un signal biologique à canaux multiples utilisant une géométrie riemannienne

Non-Patent Citations (9)

* Cited by examiner, † Cited by third party
Title
BARRY R. J.: "Characterizing pink and white noise in the human electroencephalogram", J NEURAL ENG., vol. 18, no. 3, 16 March 2021 (2021-03-16)
BARTHELEMY QUENTIN ET AL: "The Riemannian Potato Field: A Tool for Online Signal Quality Index of EEG", IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING, IEEE, USA, vol. 27, no. 2, 1 February 2019 (2019-02-01), pages 244 - 255, XP011710079, ISSN: 1534-4320, [retrieved on 20190208], DOI: 10.1109/TNSRE.2019.2893113 *
FIEDLER P. ET AL.: "Novel flexible Dry multipin electrodes for EEG: Signal quality and interfacial impedance of Ti and TiN coatings", 35TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY, 2013
LINS CALDAS A. S. ET AL.: "Towards Automatic EEG Signal Denoising by Quality Metric Optimization", INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, 2020
LIU X.: "Enhanced Auditory Steady-State Response Using an Optimized Chirp Stimulus-Evoked Paradigm", SENSORS (BASEL, vol. 19, no. 3, 12 February 2019 (2019-02-12), pages 748
NATHAN F.: "Characterizing contact impedance, signal quality and robustness as a function of the cardinality and arrangement of fingers on dry contact EEG electrodes", 36TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY, 2014
TAUTAN A. M. ET AL.: "Signal Quality in Dry Electrode EEG and the Relation to Skin-electrode Contact Impedance Magnitude", BIODEVICES2014 -INTERNATIONAL
THEA RADENTZ: "Signal Quality Evaluation of Emerging EEG Devices", FRONT. PHYSIOL., 14 February 2018 (2018-02-14)
URIGUEN J.A. ET AL., EEG ARTIFACT REMOVAL-STATE-OF-THE-ART AND GUIDELINES

Similar Documents

Publication Publication Date Title
JP4987338B2 (ja) 認知能減退を予測するためのシステムおよび方法
Fraser et al. Automated biosignal quality analysis for electromyography using a one-class support vector machine
Güler et al. Classification of EMG signals using PCA and FFT
EP3011895B1 (fr) Détermination de la charge cognitive d'un sujet à partir d'électroencéphalographie (EEG) des signaux
EP1216656B1 (fr) Procédé et dispositif d'estimation du degré de déficience neuronal du cortex cérébral
Titgemeyer et al. Can commercially available wearable EEG devices be used for diagnostic purposes? An explorative pilot study
JP7591769B2 (ja) 脳状態推定装置、コンピュータプログラム、脳状態推定方法、脳機能の検査システムおよび方法
US20110087125A1 (en) System and method for pain monitoring at the point-of-care
Yao et al. Characterizing the location and extent of myocardial infarctions with inverse ECG modeling and spatiotemporal regularization
US20180279938A1 (en) Method of diagnosing dementia and apparatus for performing the same
JP2005021176A (ja) 脳活動測定装置
CN112932486A (zh) 基于多模态的大学生情绪压力检测系统与方法
Mozaffarilegha et al. Visibility graph analysis of speech evoked auditory brainstem response in persistent developmental stuttering
Patil et al. A method for detection and reduction of stress using EEG
Johns et al. Force modelling of upper limb biomechanics using ensemble fast orthogonal search on high-density electromyography
EP4368108A1 (fr) Procédés d'obtention d'un score de qualité
Hu et al. Spectral homogeneity cross frequencies can be a quality metric for the large-scale resting EEG preprocessing
Jayas et al. Computer Aided Detection of Dominant Artifacts in Ear-EEG Signal
Qi Algorithms benchmarking for removing EOG artifacts in brain computer interface
Reddy et al. A Point-Process Approach for Tracking Valence using a Respiration Belt
Zhang et al. Research on the method of evaluating psychological stress by EEG
Alonso et al. Spatial Filter comparison for a Brain Computer Interface
CN117257303B (zh) 一种焦虑检测方法、装置、电子设备及存储介质
JP7524963B2 (ja) 情報処理装置、制御方法及びプログラム
Mumtaz et al. A study to investigate different eeg reference choices in diagnosing major depressive disorder

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN PUBLISHED

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL NO PL PT RO RS SE SI SK SM TR

RAP3 Party data changed (applicant data changed or rights of an application transferred)

Owner name: IDUN TECHNOLOGIES AG

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20241116